The neuroendocrine proteins glutamic acid decarboxylase 65 (GAD65) and IA-2 are target antigens in greater than 90 percent of individuals, who develop Type 1 diabetes but their role in induction of disease is unclear. Autoantibodies to both proteins, recognizing conformational epitopes on the surface of the native molecules, are present up to several years before clinical onset of disease, and GAD65 specific T cells may play a role in the pathogenesis of diabetes in the non-obese diabetic (NOD) mouse. High resolution mapping of conformational autoreactive epitopes in the GAD65 protein, recognized by 16 human monoclonal antibodies, together with 2- and 3-dimensional structure prediction algorithms and homology with known crystal structures, has resulted in a model of the GAD65 dimer. The autoreactive epitopes cluster in hydrophilic patches on the surface. Several epitopes have contact residues in the midst of or adjacent to the main T-cell epitopes identified for the DR*0401 haplotype, suggesting the possibility that uptake and presentation of GAD65 via surface Ig on autoreactive B cells can influence the presentation of T-cell epitopes, Preliminary experiments show that autoantibody binding can result in a 3-fold enhancement of the presentation of an adjacent epitope to a T- cell hybridoma. A 3-dimensional model has also been generated for IA-2 and highlights the possible autoreactive epitope areas specific for this molecule. The first aim proposes to use the 3-dimensional structures to complete the fine mapping of autoreactive B cell epitopes in GAD65, dissect the autoreactive epitopes in IA-2, and to study the temporal progression of epitope recognition of human disease. The second aim tests the hypothesis that GAD65 and IA-2 epitope specificity of autoreactive B-cells affects the presentation of antigen to DR4 and DQ8 restricted T-cell hybridomas. The third aim proposes to test the hypothesis that induction of strong Th1 and CTL dominated responses to GAD65, IA-2, and/or a third autoantigen, mouse insulin B-chain, can result in beta-cell destruction and diabetes in the mouse. Our long term goal is to define the pathways to chronic autoimmunity that result in pancreatic beta-cell destruction in man, identify the role of target antigens in induction and homing of autoimmunity, and eventually using the knowledge of mechanisms to design diagnostic and preventive methods.